Index: llvm/trunk/include/llvm/Transforms/Scalar/GVN.h =================================================================== --- llvm/trunk/include/llvm/Transforms/Scalar/GVN.h +++ llvm/trunk/include/llvm/Transforms/Scalar/GVN.h @@ -120,6 +120,8 @@ uint32_t lookupOrAddCall(CallInst *C); uint32_t phiTranslateImpl(const BasicBlock *BB, const BasicBlock *PhiBlock, uint32_t Num, GVN &Gvn); + bool areCallValsEqual(uint32_t Num, uint32_t NewNum, const BasicBlock *Pred, + const BasicBlock *PhiBlock, GVN &Gvn); std::pair assignExpNewValueNum(Expression &exp); bool areAllValsInBB(uint32_t num, const BasicBlock *BB, GVN &Gvn); Index: llvm/trunk/lib/Transforms/Scalar/GVN.cpp =================================================================== --- llvm/trunk/lib/Transforms/Scalar/GVN.cpp +++ llvm/trunk/lib/Transforms/Scalar/GVN.cpp @@ -1544,6 +1544,41 @@ return NewNum; } +// Return true if the value number \p Num and NewNum have equal value. +// Return false if the result is unknown. +bool GVN::ValueTable::areCallValsEqual(uint32_t Num, uint32_t NewNum, + const BasicBlock *Pred, + const BasicBlock *PhiBlock, GVN &Gvn) { + CallInst *Call = nullptr; + LeaderTableEntry *Vals = &Gvn.LeaderTable[Num]; + while (Vals) { + Call = dyn_cast(Vals->Val); + if (Call && Call->getParent() == PhiBlock) + break; + Vals = Vals->Next; + } + + if (AA->doesNotAccessMemory(Call)) + return true; + + if (!MD || !AA->onlyReadsMemory(Call)) + return false; + + MemDepResult local_dep = MD->getDependency(Call); + if (!local_dep.isNonLocal()) + return false; + + const MemoryDependenceResults::NonLocalDepInfo &deps = + MD->getNonLocalCallDependency(Call); + + // Check to see if the Call has no function local clobber. + for (unsigned i = 0; i < deps.size(); i++) { + if (deps[i].getResult().isNonFuncLocal()) + return true; + } + return false; +} + /// Translate value number \p Num using phis, so that it has the values of /// the phis in BB. uint32_t GVN::ValueTable::phiTranslateImpl(const BasicBlock *Pred, @@ -1590,8 +1625,11 @@ } } - if (uint32_t NewNum = expressionNumbering[Exp]) + if (uint32_t NewNum = expressionNumbering[Exp]) { + if (Exp.opcode == Instruction::Call && NewNum != Num) + return areCallValsEqual(Num, NewNum, Pred, PhiBlock, Gvn) ? NewNum : Num; return NewNum; + } return Num; } Index: llvm/trunk/test/Transforms/GVN/pr42605.ll =================================================================== --- llvm/trunk/test/Transforms/GVN/pr42605.ll +++ llvm/trunk/test/Transforms/GVN/pr42605.ll @@ -0,0 +1,87 @@ +; RUN: opt -gvn %s -S | FileCheck %s +; PR42605. Check phi-translate won't translate the value number of a call +; to the value of another call with clobber in between. +target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" + +@global = dso_local local_unnamed_addr global i32 0, align 4 +@.str = private unnamed_addr constant [8 x i8] c"%d, %d\0A\00", align 1 + +; Function Attrs: nofree nounwind +declare dso_local i32 @printf(i8* nocapture readonly, ...) local_unnamed_addr + +; Function Attrs: noinline norecurse nounwind readonly uwtable +define dso_local i32 @_Z3gooi(i32 %i) local_unnamed_addr #0 { +entry: + %t0 = load i32, i32* @global, align 4, !tbaa !2 + %add = add nsw i32 %t0, %i + ret i32 %add +} + +; Function Attrs: nofree nounwind uwtable +define dso_local void @noclobber() local_unnamed_addr { +entry: + %call = tail call i32 @_Z3gooi(i32 2) + %add = add nsw i32 %call, 5 + %cmp = icmp sgt i32 %add, 2 + br i1 %cmp, label %if.then, label %if.end + +if.then: ; preds = %entry + %call1 = tail call i32 @_Z3gooi(i32 3) + %add2 = add nsw i32 %call1, 5 + br label %if.end + +; Check pre happens after phitranslate. +; CHECK-LABEL: @noclobber +; CHECK: %add4.pre-phi = phi i32 [ %add2, %if.then ], [ %add, %entry ] +; CHECK: printf(i8* getelementptr inbounds {{.*}}, i32 %add4.pre-phi) + +if.end: ; preds = %if.then, %entry + %i.0 = phi i32 [ 3, %if.then ], [ 2, %entry ] + %global2.0 = phi i32 [ %add2, %if.then ], [ %add, %entry ] + %call3 = tail call i32 @_Z3gooi(i32 %i.0) + %add4 = add nsw i32 %call3, 5 + %call5 = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i64 0, i64 0), i32 %global2.0, i32 %add4) + ret void +} + +; Function Attrs: nofree nounwind uwtable +define dso_local void @hasclobber() local_unnamed_addr { +entry: + %call = tail call i32 @_Z3gooi(i32 2) + %add = add nsw i32 %call, 5 + %cmp = icmp sgt i32 %add, 2 + br i1 %cmp, label %if.then, label %if.end + +if.then: ; preds = %entry + %call1 = tail call i32 @_Z3gooi(i32 3) + %add2 = add nsw i32 %call1, 5 + br label %if.end + +; Check no pre happens. +; CHECK-LABEL: @hasclobber +; CHECK: %call3 = tail call i32 @_Z3gooi(i32 %i.0) +; CHECK-NEXT: %add4 = add nsw i32 %call3, 5 +; CHECK-NEXT: printf(i8* getelementptr inbounds ({{.*}}, i32 %global2.0, i32 %add4) + +if.end: ; preds = %if.then, %entry + %i.0 = phi i32 [ 3, %if.then ], [ 2, %entry ] + %global2.0 = phi i32 [ %add2, %if.then ], [ %add, %entry ] + store i32 5, i32* @global, align 4, !tbaa !2 + %call3 = tail call i32 @_Z3gooi(i32 %i.0) + %add4 = add nsw i32 %call3, 5 + %call5 = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([8 x i8], [8 x i8]* @.str, i64 0, i64 0), i32 %global2.0, i32 %add4) + ret void +} + +attributes #0 = { noinline norecurse nounwind readonly uwtable "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="false" "no-jump-tables"="false" "no-nans-fp-math"="false" "no-signed-zeros-fp-math"="false" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+cx8,+fxsr,+mmx,+sse,+sse2,+x87" "unsafe-fp-math"="false" "use-soft-float"="false" } + +!llvm.module.flags = !{!0} +!llvm.ident = !{!1} + +!0 = !{i32 1, !"wchar_size", i32 4} +!1 = !{!"clang version 10.0.0 (trunk 369798)"} +!2 = !{!3, !3, i64 0} +!3 = !{!"int", !4, i64 0} +!4 = !{!"omnipotent char", !5, i64 0} +!5 = !{!"Simple C++ TBAA"}